New Pi Stratum 1 Check

#1

Hi all setup a pi stratum 1 server following the guide here
https://www.ntpsec.org/white-papers/stratum-1-microserver-howto/

Using the adafruit ultimate HAT board. Just checking my stats if everything is working as it should as i am new to this. Thanks

X0Gh3BV.png837x215 11.3 KB

Wondering if this looks right? Concerned it seems the GPS has been flagged as a false ticker however the PPS seems to be working.

If you could also check the ntpviz info at time.arihanc.com and see if it looks reasonable that would be appreciated.

Thanks

#2

Looks like you still need a round of calibration - to calculate the value to use to adjust your ‘time1’ parameter for the latency in the connection to your refclock.

See http://www.catb.org/gpsd/gpsd-time-service-howto.html#_ntp_performance_tuning

In a nutshell:

  • Set your SHM0 and SHM1 to be ignored for sync purposes (‘noselect’ clause)
  • Set some high-quality time sources as servers
  • Enable peerstats and/or loopstats logging
  • Run for a few hours (probably at least four?)
  • Calculate the average of the offset between your SHM0 and the time from those high-quality sources
  • Update the ‘time1’ parameter for SHM0 to the negative of that average
#3

Many GPS devices have the same problem of delayed GPS time output, including my Garmin 18x OEM. I am too lazy to calculate the offset of SHM(0), so I disable it and get second data from network then calibrate it by SHM(1) PPS data. Of course this configuration cannot work in isolated network, but should not be a problem for NTP pool servers.

My current ntpq -p output:

     remote           refid      st t when poll reach   delay   offset  jitter
*SHM(1)          .PPS.            0 l   33   64  377    0.000    0.000   0.004
-118-163-81-62.H 192.168.0.3      2 u   17   64  377    4.036    1.621   0.361
-211-22-103-157. 192.168.0.3      2 u   58   64  377    4.240    1.612   0.111
-118-163-81-61.H 192.168.0.2      2 u   23   64  377    4.163    1.539   0.164
+211-22-103-158. 192.168.0.2      2 u   56   64  377    4.468    0.947   0.625
-118-163-81-63.H 192.168.0.2      2 u   61   64  377    4.044    1.918   0.385
-223.255.185.2   .192..           1 u   20   64  357   54.009  -17.774   0.526
+ntp-a3.nict.go. .NICT.           1 u   16   64  257   33.138    0.017   0.161
#4

Hi thanks for the replies.
@tychotithonus Am i correct in saying i can use the offset generated by ntpviz?
http://time.arihanc.com/week/index.html
looking at the weekly stats for SHM(0) it seems to be median = -0.515241 so id put it as a positive correct?

@alica Thats handy to know thanks, although id like to get the GPS working if i can its not the end of the world just to use PPS.

#5

@arihancc Yep, that looks about right. If you want to be extra precise, you might look at the actual peerstats file data to see if there are any major outliers that should be discarded, create a cleaned up copy of the peerstats, and calculate the average on that clean copy using the awk snippet.